NUREG/CR-6427 "Assessment of the DCH Issue for Plants with Ice Condenser Containments"

This report (NUREG/CR-6427) addresses the Direct Containment Heating (DCH) issue for all Westinghouse plants with ice condenser containments. There are ten operating ice condenser plants located at five sites in the U.S. DCH phenomena in ice condenser plants are different in some important aspects from DCH phenomena in other Pressurized Water Reactors (PWRs) in that they have ice beds to suppress Design Basis Accident (DBA) steam loads, AC-powered igniters to control hydrogen concentrations in the atmosphere, small containment volumes, and containment buildings with low ultimate capacities to withstand internal pressures. Unlike PWRs with large dry or subatmospheric containments, the DCH issue for ice condenser plants could not be resolved by a probabilistic comparison of containment loads versus containment strength. The approach taken here is to provide an expansion of a probabilistic framework, which represents a simplification of the NUREG-1150 containment event tree for Sequoyah. The containment event tree is intended to give each containment challenge its proper probabilistic weighting based on plant specific core damage frequencies, phenomenological probabilities, and plant specific fragility curves. The probabilistic framework addresses DCH induced overpressure failures in the context of all significant early containment failure modes. These include DCH overpressure failures, thermal failures of the containment liner, non-DCH hydrogen combustion overpressure failures, and non-explosive steam spike overpressure failures. The most significant finding of this study was that the early containment failure probability is dominated by non-DCH hydrogen combustion events rather than DCH events. This is because the HPME probability is small, the SBO probabilities are small, and because containment loads in non-station blackouts are not containment threatening. The CONTAIN code was used exclusively to calculate containment, loads resulting from DCH, non-DCH hydrogen combustion, and non-explosive steam spikes for representative station blackout and non-station blackout scenarios. CONTAIN calculations show that no ice condenser plant is inherently robust to all credible DCH or hydrogen combustion events in station blackouts. CONTAIN predictions show that containment loads in nonstation blackout are not containment threatening for any reasonable plant damage state. Consistent with perceptions of the technical community, this study shows that ice condenser plants are substantially more sensitive to early containment failure than PWRs with large dry or subatmospheric containments. A plant-specific evaluation of the containment event tree showed that all plants, except McGuire, have an early failure probability within the range 0.35% to 5.8% for full power internal events. The early containment failure probability was 13.9% for McGuire. The higher containment failure probability is dominated by the relatively higher station blackout probability and relatively weaker containment for McGuire. Reduction in the probability of a struck open power-operated relief valve (PORV) after uncovery of the top of active fuel (UTAF) had no signification impact on the conclusions of this study. Reduction in the hot leg failure probability increases the probability of early containment failure for those plants with a large SBO frequency, but not to the point that conclusions regarding compliance with NRC goals would change. An additional sensitivity study assuming